Search NASA⌕ Search

SEARCH · Search NASA

Results for “loss function regularization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3

Estimating conditional quantiles with the help of the pinball loss

Using the so-called pinball loss for estimating conditional quantiles is a well-known tool in both statistics and machine learning. So far, however, only little work has been done to quantify the efficiency of this tool for non-parametric (modified) empirical risk minimization approaches. The goal of this work is to fill this gap by establishing inequalities that describe how close approximate pinball risk minimizers are to the corresponding conditional quantile. These inequalities, which hold under mild assumptions on the data-generating distribution, are then used to establish so-called variance bounds which recently turned out to play an important role in the statistical analysis of (modified) empirical risk minimization approaches. To illustrate the use of the established inequalities, we then use them to establish an oracle inequality for support vector machines that use the pinball loss. Here, it turns out that we obtain learning rates which are optimal in a min-max sense under some standard assumptions on the regularity of the conditional quantile function.

97 MATHEMATICS AND COMPUTING↗

Alignment of morphology during high spatial frequency periodic structure formation in GaAs

The interaction among multiple intense ultrashort laser pulses and solids is known to provide a regular nanoscale surface corrugation. A coupled mechanism has been identified that operates in a specific range of fluences in GaAs that exhibits transient loss of the imaginary part of the dielectric function and Χ 2 , which produces a unique corrugation known as high spatial frequency laser induced periodic surface structures (HSFL). The final structures have 180 nm periods, and their alignment perpendicular to the laser polarization is first observed in an intermediate morphology with correlation distances of 150±40 nm. Quantum molecular dynamics simulations suggest that HSFL self-assembly is initiated when the intense laser field softens the interatomic binding potential, which leads to an ultrafast generation of point defects. The morphological evolution begins as self-interstitial diffusion, driven by stress relaxation, to the surface producing 1–2 nm tall islands. An ab initio calculation of excited electron concentration combined with a Drude-Lorentz model of the excited GaAs dielectric function is used to determine that the conditions for SPP coupling at HSFL formation fluences are both satisfied and occur at wavelengths that are imprinted into the observed surface morphologies. The evolution of these morphologies is reported on as the interplay between surface plasmon polaritons that localize defect generation within the structures present on the previous laser exposure and stress relaxation driven defect diffusion.

36 MATERIALS SCIENCE↗

Multidimensional Distributional Neural Network Output Demonstrated in Super‐Resolution of Surface Wind Speed

Accurate quantification of uncertainty in neural network predictions remains a central challenge for scientific applications involving high-dimensional, correlated data. While existing methods capture either aleatoric or epistemic uncertainty, few offer closed-form, multidimensional distributions that preserve spatial correlation while remaining computationally tractable. In this work, we present a framework for training neural networks with a multidimensional Gaussian loss, generating a closed-form predictive distribution over outputs informed by non-identically distributed training data. Our approach captures aleatoric uncertainty by iteratively estimating the means and covariance matrices, and is demonstrated on a super-resolution example out-of-training-sample. We leverage a Fourier representation of the covariance matrix to stabilize network training and preserve spatial correlation. We introduce a novel regularization strategy—referred to as information sharing—that interpolates between image-specific and global covariance estimates, enabling convergence of the super-resolution downscaling network trained on image-specific distributional loss functions. This framework allows for efficient sampling, explicit correlation modeling, and extensions to more complex distribution families all without disrupting prediction performance. We demonstrate the method on a surface wind speed downscaling task and discuss its broader applicability to uncertainty-aware prediction in scientific models.

17 WIND ENERGY↗

FunDiff: diffusion models over function spaces for physics-informed generative modeling

Recent advances in generative modeling-particularly diffusion models and flow matching-have been widely used for synthesizing discrete data such as images and videos. However, adapting these models to physical applications remains challenging, as the quantities of interest are continuous functions governed by complex physical laws. To address this, we introduce FunDiff, an efficient and robust framework for generative modeling in function spaces. FunDiff combines a latent diffusion process with a function autoencoder architecture to handle input functions with varying discretizations, generates continuous functions that can be evaluated at arbitrary locations, and seamlessly incorporate physical priors. These priors are enforced through architectural constraints or physics-informed loss functions, ensuring that generated samples satisfy fundamental physical laws. We theoretically establish minimax optimality guarantees for density estimation in function spaces, demonstrating that diffusion-based estimators achieve optimal convergence rates under suitable regularity conditions. We further demonstrate the practical effectiveness of FunDiff across diverse applications in fluid dynamics and solid mechanics. Empirical results indicate that our method can generate physically consistent samples with high fidelity to the target distribution, and exhibit robustness to noisy and low-resolution data.

Wang, Sifan [Yale University, New Haven, CT (Unite↗

Alpha image reconstruction (AIR): A new iterative CT image reconstruction approach using voxel-wise alpha blending

Purpose: Iterative image reconstruction gains more and more interest in clinical routine, as it promises to reduce image noise (and thereby patient dose), to reduce artifacts, or to improve spatial resolution. Among vendors and researchers, however, there is no consensus of how to best achieve these aims. The general approach is to incorporatea priori knowledge into iterative image reconstruction, for example, by adding additional constraints to the cost function, which penalize variations between neighboring voxels. However, this approach to regularization in general poses a resolution noise trade-off because the stronger the regularization, and thus the noise reduction, the stronger the loss of spatial resolution and thus loss of anatomical detail. The authors propose a method which tries to improve this trade-off. The proposed reconstruction algorithm is called alpha image reconstruction (AIR). One starts with generating basis images, which emphasize certain desired image properties, like high resolution or low noise. The AIR algorithm reconstructs voxel-specific weighting coefficients that are applied to combine the basis images. By combining the desired properties of each basis image, one can generate an image with lower noise and maintained high contrast resolution thus improving the resolution noise trade-off. Methods: All simulations and reconstructions are performed in native fan-beam geometry. A water phantom with resolution bar patterns and low contrast disks is simulated. A filtered backprojection (FBP) reconstruction with a Ram-Lak kernel is used as a reference reconstruction. The results of AIR are compared against the FBP results and against a penalized weighted least squares reconstruction which uses total variation as regularization. The simulations are based on the geometry of the Siemens Somatom Definition Flash scanner. To quantitatively assess image quality, the authors analyze line profiles through resolution patterns to define a contrast factor for contrast-resolution plots. Furthermore, the authors calculate the contrast-to-noise ratio with the low contrast disks and the authors compare the agreement of the reconstructions with the ground truth by calculating the normalized cross-correlation and the root-mean-square deviation. To evaluate the clinical performance of the proposed method, the authors reconstruct patient data acquired with a Somatom Definition Flash dual source CT scanner (Siemens Healthcare, Forchheim, Germany). Results: The results of the simulation study show that among the compared algorithms AIR achieves the highest resolution and the highest agreement with the ground truth. Compared to the reference FBP reconstruction AIR is able to reduce the relative pixel noise by up to 50% and at the same time achieve a higher resolution by maintaining the edge information from the basis images. These results can be confirmed with the patient data. Conclusions: To evaluate the AIR algorithm simulated and measured patient data of a state-of-the-art clinical CT system were processed. It is shown, that generating CT images through the reconstruction of weighting coefficients has the potential to improve the resolution noise trade-off and thus to improve the dose usage in clinical CT.

60 APPLIED LIFE SCIENCES↗

Stationary bubbles and their tunneling channels toward trivial geometry

In the path integral approach, one has to sum over all histories that start from the same initial condition in order to obtain the final condition as a superposition of histories. Applying this into black hole dynamics, we consider stable and unstable stationary bubbles as a reasonable and regular initial condition. We find examples where the bubble can either form a black hole or tunnel toward a trivial geometry, i.e., with no singularity nor event horizon. We investigate the dynamics and tunneling channels of true vacuum bubbles for various tensions. In particular, in line with the idea of superposition of geometries, we build a classically stable stationary thin-shell solution in a Minkowski background where its fate is probabilistically given by non-perturbative effects. Since there exists a tunneling channel toward a trivial geometry in the entire path integral, the entire information is encoded in the wave function. This demonstrates that the unitarity is preserved and there is no loss of information when viewed from the entire wave function of the universe, whereas a semi-classical observer, who can see only a definitive geometry, would find an effective loss of information. Ultimately, this may provide a resolution to the information loss dilemma.

79 ASTRONOMY AND ASTROPHYSICS↗

Drain Tank Information for Developing Design Basis of the Preliminary Design - R00

Tokamak Cooling Water System (TCWS) drain tanks (DTs) serve two functions: normal operation and safety operation. Normal DTs are used for regular maintenance operations when draining is necessary. Safety DTs are used to receive the water leaked into the Vacuum Vessel (VV) after an in-vessel loss of cooling accident (LOCA) event. The preliminary design of the DTs shall be based on the information provided by this document. The capacity of the normal DTs is estimated based on the internal volume of in-vessel components [e.g., First Wall/Blanket (FW/BLK) and Divertor (DIV)], Neutral Beam Injector (NBI) components, and TCWS piping, heat exchangers, electric heaters, pump casing, pressurizers, and valves. Water volumes have been updated based on 2004 design information, changes adopted because of approved Project Change Requests (PCRs), and data verification by U.S. ITER. Two tanks will store water from normal draining operations of the FW/BLK and DIV Primary Heat Transfer Systems (PHTSs). One tank will store water from normal draining operations of the NBI PHTS. The capacity of the safety DTs is based on analysis of a design-basis accident:1 a large leak from in-vessel components. There are two safety DTs that will receive water from a VV LOCA event and drainage from the VV, as needed. In addition, there is one sump tank for the DIV that will be used for collecting drain water from the draining and drying processes and specifically for draining the DIV system as the DIV cassettes lines are at a lower elevation than the DT connection point. Information documented in this report must be refined and verified during the preliminary design of the DTs, and there are several aspects to be considered to complete the preliminary design. Input to these design considerations is discussed in this report and includes, but is not limited to, water inventory; operating procedures/maintenance; Failure Modes and Effects Analysis (FMEA); tank layout anddimensions, including design margin; classification under French Nuclear Pressure Directives, Equipements Sous Pression Nucleaires (ESPN); and adaptations for construction.

Cooling Water↗

Drain Tank Information for Developing Design Basis of the Preliminary Design

Tokamak Cooling Water System (TCWS) drain tanks (DTs) serve two functions: normal operation and safety operation. Normal DTs are used for regular maintenance operations when draining is necessary. Safety DTs are used to receive the water leaked into the Vacuum Vessel (VV) after an in-vessel loss of coolant accident (LOCA) event. The preliminary design of the DTs shall be based on the information provided by this document. The capacity of the normal DTs is estimated based on the internal volume of in-vessel components [e.g., First Wall/Blanket (FW/BLK) and Divertor (DIV)]; Neutral Beam Injector (NBI) components; and TCWS piping, heat exchangers, electric heaters, pump casing, pressurizers, and valves. Water volumes have been updated based on 2004-design information, changes adopted because of approved Project Change Requests (PCRs), and data verification by US ITER and AREVA Federal Services, the US ITER A and E Company. Two tanks will store water from normal draining operations of the FW/BLK and DIV Primary Heat Transfer Systems (PHTSs). One tank will store water from normal draining operations of the NBI PHTS. The capacity of the safety DTs is based on analysis of a design basis accident: a large leak from in-vessel components. There are two safety DTs that will receive water from a VV LOCA event and drainage from the VV, as needed. In addition, there is one sump tank for the DIV that will be used for collecting drain water from the draining and drying processes and specifically for draining the DIV system as the DIV cassette lines are at a lower elevation than the DT connection point. Information documented in this report must be refined and verified during the preliminary design of the DTs, and there are several aspects to be considered to complete the preliminary design. Input to these design considerations is discussed in this report and includes, but is not limited to, water inventory; operating procedures/maintenance; Failure Modes and Effects Analysis (FMEA); tank layout and dimensions, including design margin; classification under French Nuclear Pressure Directives, Equipements Sous Pression Nucleaires (ESPN); and adaptations for construction.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Recipes for when physics fails: recovering robust learning of physics informed neural networks

Abstract Physics-informed neural networks (PINNs) have been shown to be effective in solving partial differential equations by capturing the physics induced constraints as a part of the training loss function. This paper shows that a PINN can be sensitive to errors in training data and overfit itself in dynamically propagating these errors over the domain of the solution of the PDE. It also shows how physical regularizations based on continuity criteria and conservation laws fail to address this issue and rather introduce problems of their own causing the deep network to converge to a physics-obeying local minimum instead of the global minimum. We introduce Gaussian process (GP) based smoothing that recovers the performance of a PINN and promises a robust architecture against noise/errors in measurements. Additionally, we illustrate an inexpensive method of quantifying the evolution of uncertainty based on the variance estimation of GPs on boundary data. Robust PINN performance is also shown to be achievable by choice of sparse sets of inducing points based on sparsely induced GPs. We demonstrate the performance of our proposed methods and compare the results from existing benchmark models in literature for time-dependent Schrödinger and Burgers’ equations.

97 MATHEMATICS AND COMPUTING↗

Study of the Vertical Magnetic Field in Face-on Galaxies Using Faraday Tomography

Faraday tomography allows astronomers to probe the distribution of the magnetic field along the line of sight (LOS), but that can be achieved only after the Faraday spectrum is interpreted. However, the interpretation is not straightforward, mainly because the Faraday spectrum is complicated due to a turbulent magnetic field; it ruins the one-to-one relation between the Faraday depth and the physical depth, and appears as many small-scale features in the Faraday spectrum. In this paper, by employing “simple toy models” for the magnetic field, we describe numerically as well as analytically the characteristic properties of the Faraday spectrum. We show that the Faraday spectrum along “multiple LOSs” can be used to extract the global properties of the magnetic field. Specifically, considering face-on spiral galaxies and modeling turbulent magnetic field as a random field with a single coherence length, we numerically calculate the Faraday spectrum along a number of LOSs and its shape-characterizing parameters, that is, the moments. When multiple LOSs cover a region of ≳(10 coherence length){sup 2}, the shape of the Faraday spectrum becomes smooth and the shape-characterizing parameters are well specified. With the Faraday spectrum constructed as a sum of Gaussian functions with different means and variances, we analytically show that the parameters are expressed in terms of the regular and turbulent components of the LOS magnetic field and the coherence length. We also consider the turbulent magnetic field modeled with a power-law spectrum, and study how the magnetic field is revealed in the Faraday spectrum. Our work suggests a way to obtain information on the magnetic field from a Faraday tomography study.

79 ASTRONOMY AND ASTROPHYSICS↗

Test Beam Results of 3D Silicon Pixel Sensors for the ATLAS upgrade

Results on beam tests of 3D silicon pixel sensors aimed at the ATLAS Insertable-B-Layer and High Luminosity LHC (HL-LHC) upgrades are presented. Measurements include charge collection, tracking efficiency and charge sharing between pixel cells, as a function of track incident angle, and were performed with and without a 1.6 T magnetic field oriented as the ATLAS Inner Detector solenoid field. Sensors were bump bonded to the front-end chip currently used in the ATLAS pixel detector. Full 3D sensors, with electrodes penetrating through the entire wafer thickness and active edge, and double-sided 3D sensors with partially overlapping bias and read-out electrodes were tested and showed comparable performance. Full and partial 3D pixel detectors have been tested, with and without a 1.6T magnetic field, in high energy pion beams at the CERN SPS North Area in 2009. Sensors characteristics have been measured as a function of the beam incident angle and compared to a regular planar pixel device. Overall full and partial 3D devices have similar behavior. Magnetic field has no sizeable effect on 3D performances. Due to electrode inefficiency 3D devices exhibit some loss of tracking efficiency for normal incident tracks but recover full efficiency with tilted tracks. As expected due to the electric field configuration 3D sensors have little charge sharing between cells.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

A comparison between bright points in a coronal hole and a quiet-sun region

A comparison is made of the morphological structure and temporal behavior of the emission from coronal bright points in a coronal hole and a quiet region, using data from the Harvard EUV experiment on Skylab. It is found that, in both regions, coronal bright points are located at network boundaries and cover a range of sizes from 10 to 40 in in linear extent. In a given bright pint, the peaks of emission in the six different lines, measured simultaneously through the same instrument slit, are not always cospatial, implying that bright points consist of a complex of small-scale loops at different temperatures. The intensity of bright points in both regions is also characterized by a significant temporal variability in all the wavelengths measured. This variability exhibits no regular periodicity. Yet the ratio of the varying (ac) to the constant (dc) components of the emission, in all the bright points studied, has a local maximum at 1-2 x 10 to the 5th k which coincides with the peak of the radiative loss function, and another local maximum at Mg x (1.4 x 10 to the 6th K). It is found that coronal bright points in a coronal hole or a quiet region are indistinguishable structures, and, therefore, conclude that they are independent of the overlying background corona.

Habbal, Shadia Rifai↗

Linking Changes in Epithelial Morphogenesis to Cancer Mutations Using Computational Modeling

Most tumors arise from epithelial tissues, such as mammary glands and lobules, and their initiation is associated with the disruption of a finely defined epithelial architecture. Progression from intraductal to invasive tumors is related to genetic mutations that occur at a subcellular level but manifest themselves as functional and morphological changes at the cellular and tissue scales, respectively. Elevated proliferation and loss of epithelial polarization are the two most noticeable changes in cell phenotypes during this process. As a result, many three-dimensional cultures of tumorigenic clones show highly aberrant morphologies when compared to regular epithelial monolayers enclosing the hollow lumen (acini). In order to shed light on phenotypic changes associated with tumor cells, we applied the bio-mechanical IBCell model of normal epithelial morphogenesis quantitatively matched to data acquired from the non-tumorigenic human mammary cell line, MCF10A. We then used a high-throughput simulation study to reveal how modifications in model parameters influence changes in the simulated architecture. Three parameters have been considered in our study, which define cell sensitivity to proliferative, apoptotic and cell-ECM adhesive cues. By mapping experimental morphologies of four MCF10A-derived cell lines carrying different oncogenic mutations onto the model parameter space, we identified changes in cellular processes potentially underlying structural modifications of these mutants. As a case study, we focused on MCF10A cells expressing an oncogenic mutant HER2-YVMA to quantitatively assess changes in cell doubling time, cell apoptotic rate, and cell sensitivity to ECM accumulation when compared to the parental non-tumorigenic cell line. By mapping in vitro mutant morphologies onto in silico ones we have generated a means of linking the morphological and molecular scales via computational modeling. Thus, IBCell in combination with 3D acini cultures can form a computational/experimental platform for suggesting the relationship between the histopathology of neoplastic lesions and their underlying molecular defects.

59 BASIC BIOLOGICAL SCIENCES↗

Multidisciplinary design analysis and optimisation of a reference offshore wind plant

Motivated by the need to develop reference wind energy systems for optimisation and technology assessment studies, the International Energy Agency Wind Task 37 on Wind Energy Systems Engineering is developing a reference offshore wind power plant at the Dutch offshore wind energy areas Borssele III and IV. This work presents a comparison between two approaches for developing the preliminary design of an offshore wind plant turbine layout, electrical collection system, and support structures. The first is a sequential approach, where components of the wind farm are optimised sequentially, each with its own objective function, thus neglecting potential interactions between them. The second approach uses Multidisciplinary Design Analysis and Optimisation (MDAO), where all components are jointly optimised with the overall system levelised cost of energy (LCOE) as a global objective function. Studying the cases of regular and irregular layouts, the integrated approach always shows a greater improvement in the LCOE of the final design compared to the design resulting from the traditional sequential approach. The most significant trade-off exploited by the MDAO approach used in this study is between losses in energy production due to turbine wake effects and the costs of electrical cable infrastructure.

17 WIND ENERGY↗

A NEW DIFFERENTIAL AND ERRANT BEAM CURRENT MONITOR FOR THE SNS* ACCELERATOR

A new Differential and errant Beam Current Monitor (DBCM) is being implemented for both the Spallation Neutron Source's Medium Energy Beam Transport (MEBT) and the Super Conducting Linac (SCL) accelerator sections. These new current monitors will abort the beam when the difference between two toroidal pickups exceeds a threshold. The MEBT DBCM will protect the MEBT chopper target, while the SCL DBCM will abort beam to minimize fast beam losses in the SCL cavities. The new DBCM will also record instances of errant beam, such as beam dropouts, to assist in further optimization of the SNS Accelerator. A software Errant Beam Monitor was implemented on the regular BCM hardware to study errant beam pulses. The new system will take over this functionality and will also be able to abort beam on pulse-to-pulse variations. Because the system is based on the FlexRIO hardware and programmed in LabVIEW FPGA, it will be able to abort beam in about 5 us. This paper describes the development, implementation, and initial test results of the DBCM, as well as errant beam examples.

beam current monitor↗

How Well Does Kohn–Sham Regularizer Work for Weakly Correlated Systems?

Kohn–Sham regularizer (KSR) is a differentiable machine learning approach to finding the exchange-correlation functional in Kohn–Sham density functional theory that works for strongly correlated systems. Here we test KSR for a weak correlation. We propose spin-adapted KSR (sKSR) with trainable local, semilocal, and nonlocal approximations found by minimizing density and total energy loss. We assess the atoms-to-molecules generalizability by training on one-dimensional (1D) H, He, Li, Be, and Be 2+ and testing on 1D hydrogen chains, LiH, BeH 2 , and helium hydride complexes. The generalization error from our semilocal approximation is comparable to other differentiable approaches, but our nonlocal functional outperforms any existing machine learning functionals, predicting ground-state energies of test systems with a mean absolute error of 2.7 mH.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Gravitational self-force on a particle in eccentric orbit around a Schwarzschild black hole

We present a numerical code for calculating the local gravitational self-force acting on a pointlike particle in a generic (bound) geodesic orbit around a Schwarzschild black hole. The calculation is carried out in the Lorenz gauge: For a given geodesic orbit, we decompose the Lorenz-gauge metric perturbation equations (sourced by the delta-function particle) into tensorial harmonics, and solve for each harmonic using numerical evolution in the time domain (in 1+1 dimensions). The physical self-force along the orbit is then obtained via mode-sum regularization. The total self-force contains a dissipative piece as well as a conservative piece, and we describe a simple method for disentangling these two pieces in a time-domain framework. The dissipative component is responsible for the loss of orbital energy and angular momentum through gravitational radiation; as a test of our code we demonstrate that the work done by the dissipative component of the computed force is precisely balanced by the asymptotic fluxes of energy and angular momentum, which we extract independently from the wave-zone numerical solutions. The conservative piece of the self-force does not affect the time-averaged rate of energy and angular-momentum loss, but it influences the evolution of the orbital phases; this piece is calculated here for the first time in eccentric strong-field orbits. As a first concrete application of our code we recently reported the value of the shift in the location and frequency of the innermost stable circular orbit due to the conservative self-force [Phys. Rev. Lett. 102, 191101 (2009)]. Here we provide full details of this analysis, and discuss future applications.

79 ASTRONOMY AND ASTROPHYSICS↗

AADL: Anderson Accelerated Deep Learning

We propose a stable, distributed approach to perform AA that accelerates the convergence rate of stochastic first-order optimizers to train neural networks. Differently from previous works, we do not alter neither the scheme to perform AA nor the loss function minimized during the training. To improve robustness against stagnation, we customize general guidelines that suggest to relax the frequency of AA corrections by performing AA only at the end of an entire training epoch. To improve robustness of AA against the stochastic oscillations of first-order optimizers, we average the gradients computed on consecutive stochastic optimization updates. The improved regularity of the converging sequence and the reduced amplitude of stochastic oscillations across consecutive optimization steps allows AA to efficiently extrapolate an improved converging sequence, thereby overcoming limitations of existing approaches to perform AA on stochastic optimization.

Lupo Pasini, Massimiliano [Oak Ridge National Lab.↗